A time domain reconstruction method of randomly sampled frequency sparse signal

In this paper stochastic sampling as a method of frequency sparse signal acquisition is presented. Basic principle of compressed sensing is reviewed, with emphasis on nonuniform sampling and signal reconstruction methods. A robust time domain reconstruction method of randomly sampled signal through compressed sensing approach is proposed. The presented reconstruction algorithm is evaluated by means of simulations, with comparison to conventional compressed sensing reconstruction and the most common practical issues taken into account. Simulation results indicate that the proposed reconstruction method is resistent to high levels of quantization and uncorrelated noise. Experiments with real hardware were also performed, results of which confirm the ability of stochastic sampling framework to overcome the Nyquist limit of analog-to-digital converters.